Stall Speed Bug?

[dtango]

Originally posted by Tilt
Whilst I realise your initially testing clean to compare with RL figures......could you not test under AH auto speed setting for power on and power off

You should be able to obtain very accurate comparison between AH's power on and power off stall points then.................

Tilt - the above tests were done by manually flying, no auto-speed involved.


Tango, XO
412th FS Braunco Mustangs

[hitech]

Widewing: Your test is not really measuring power on stall. The stall occurs as soon as you see the cockpit buffet. Also you will need to provide control input both roll and yaw to maintain level flight. Other wise you are just measuring our trim settings. Power on stall is very hard to fly and test do to the fact you have to get way behind the power curve before you start adding throttle.


HiTech

[Widewing]

Originally posted by hitech
Widewing: Your test is not really measuring power on stall. The stall occurs as soon as you see the cockpit buffet. Also you will need to provide control input both roll and yaw to maintain level flight. Other wise you are just measuring our trim settings. Power on stall is very hard to fly and test do to the fact you have to get way behind the power curve before you start adding throttle.


HiTech

Thanks for replying HiTech. I appreciate your interest.

If you are talking about stall buffet, then this is an impending stall, or wing root stall, not yet fully stalled. This assumes that the cockpit shake represents stall buffet. This also assumes that your code stalls the wing progressively from root to tip.

I've flown several power-on stalls in the Grumman C-1A. Not especially difficult to do. The S-2 and C-1 gave almost no warning of a stall, thus a stall-shaker was installed on the yoke. This worked via an airflow sensor installed on the upper surface of the outboard port wing. If the shaker began to vibrate it told you that that you were approximately 5 knots above stall.

I've flown the the identical profiles described previously with combat trim off and saw no appreciable difference in stall speed. My results were no different; the planes stall at higher speeds power-on than with power-off.

Note that flying with a nose-high profile reduces shake duration because speed is bleeding somewhat faster than when nearly level.

Even using cockpit shake as the reference and trimming manually, the discrepancy is still there. Power-on stall occurs about 12 mph higher than power-off stall. In the real F4U and F6F, the opposite occurs.

F4U-1A, clean.
Power-on stall to shake (44" MAP @ 2550 rpm) 107 mph
Power-off stall to shake (0" MAP @ 2550 rpm) 95 mph

F6F-5, clean.
Power-on stall to shake (44" MAP @ 2550 rpm) 98 mph
Power-off stall to shake (0" MAP @ 2550 rpm) 86 mph

Here's the films. By the way, if the aircraft seems to bank to one side while climbing, this was due to not having a visual reference of the horizon, not to excessive torque.

F6F-5 stalls, manual trim

F4U-1A stalls, manual trim

Perhaps Mace can offer his opinion. Mace is a graduate of the Naval Test Pilot School at Patuxent River.

Most folks probably haven't seen NACA Technical Report No. 829. Use this link to download a pdf copy from NASA's technical report server. The report is titled: Summary of measurements in Langley full-scale tunnel of maximum lift coefficients and stalling characteristics of airplanes. Planes evaluated were the F6F-3, F4U-1, P-51B and P-63A, among several others.

My regards,

Widewing

[Mace2004]

Originally posted by hitech
Widewing: Your test is not really measuring power on stall. The stall occurs as soon as you see the cockpit buffet. Also you will need to provide control input both roll and yaw to maintain level flight. Other wise you are just measuring our trim settings. Power on stall is very hard to fly and test do to the fact you have to get way behind the power curve before you start adding throttle.


HiTech

HiTech,

Who defines the initial onset of buffet as the stall point?  Initial buffet is usually caused just by localized flow separation often due to a non-uniformity in the wing such as a stall strip, dog tooth, gun ports, etc.  It can also be caused by changing flow patterns over the wing.  Some wings will go deep into buffet before losing significant lift but of course some wings will lose lift immediately after or coincidentally with onset of buffet; however this would not be a suitable fighter aircraft wing. I suppose that the test conditions could define stall as the initial onset of buffet if such is the case (or an ultraconservative number is desired such as GA or transport aircraft) but I've never seen stall defined this way.  I also don't think I'd want to fight an airplane with such a characteristic.

Also, if you define power-on stall as initial onset of buffet Widewing's numbers would be even worse (higher) and nowhere near Grumman's.

In every case I've ever seen stall is either the point that the nose or wing drops or the minimum usable flying speed.  Minimum usable flying speed is identified as a result of some characteristic such as loss of control on one or more axis or high rate of descent (mushing).  A "mushing" stall is very subjective, susceptible to pilot perception, and difficult to repeat since there is no definitive stall point, that's why a uniform, objective value of ROD or specific AOA can be defined.  If the limit is minimum usable flying speed that becomes the defined stall.

I agree that control input should be made; however, the preferred control would be rudder followed by the minimum amount of aileron (because aileron inputs can induce tip stall).

In any case and regardless of the test procedure I see no way that the power-on and power-off stall speeds could possibly end up reversed per the actual flight-test data which supports Widewing's arguement that something is amiss.

Edit:  I just finished going over the NACA report and it's very interesting stuff.  They never address the effect of torque but it does look like the left roll-off for the F4U is significantly driven by stall induced by increased local AOA resulting from the prop.  They added a stall strip to the right leading edge but it's not clear if this is addition to the stall strip installed in later production models.  From what I can see it appears significantly larger than what I recall on the production planes.

Mace

[Mace2004]

Originally posted by gripen
There should be no large difference wether the stall is generated during level flight or in glide, the later is just easier to reach and measure.

AFAIK the speed is directly CAS in the AH.

gripen

I must not have been clear.  I was comparing the difference in nose-high (climbing) stalls and level flightstalls.  Many (probably most) aircraft will either climb or descend a bit just prior to stall when doing the level test technique but the differences in recorded speed is very minimum.  I agree that a level test compared to a descending test would give essentially the same results.  

That said, a climbing stall will result in a more rapid reduction of airspeed which will skew the numbers downward.  This, coupled with the lift due to thrust (the helicopter effect) will result in artificially low stall speeds.  Great if you're a contractor trying to meet the stall requirements of a government contract but sucks when you're the pilot and find you stall five knots above the posted stall speed on approach.

As far as the speed in film viewer it appears to clearly be true airspeed.  The E6B correctly shows identical true and indicated speeds on the deck but much higher true at altitude which also agrees with the film viewer speeds.

Mace

[dtango]

Mace:

well since HT wrote the code he get's to define what exceeding aircraft clmax looks like so I take it from his comment that in AH physics cockpit buffet indicates that a stall has occured .

Also, I'm certain Pyro and HT are aware of the NACA report 829 that Widewing is referring to.  We've used it in the past to discuss other clmax data of particular aircraft.  If you look at that doc it discusses a variety of things including the concept of initial stall vs. the progression of the stall over the wing.  I know back when AH2 was coming out I had some conversations with Pyro that they were going to include more force vectors along the wings in AH2 which would enable finer modeling of lift across the wings.  Back then I inferred from Pyro's comments that one of the things this would allow them to do was to model stall progression across the wing among other things (e.g. increased lift due to propwash etc. etc.) since Cl doesn't just drop to 0 when Clmax is exceeded, and also that Clmax may be exceeded on certain parts of the wing prior to others.

******
HT/Pyro:

I'm curious as to what you meant by "getting behind the power curve" in relation to performing a power-on stall in AH.  I'm guessing you mean the power-required curve (drag) but not sure exactly what you were getting at in your statement.  Is there a way that we should be conducting power-on stall tests for AH planes?  I'm with Widewing that I would think power-on stalls should occur at lower airspeeds vs. power-off stalls.

Tango, XO
412th FS Braunco Mustangs

[gripen]

Originally posted by Mace2004

As far as the speed in film viewer it appears to clearly be true airspeed.  The E6B correctly shows identical true and indicated speeds on the deck but much higher true at altitude which also agrees with the film viewer speeds.

I mean that the AH speeds (from cockpit or film viewer) should be without any error caused by AoA. This means that the IAS from cockpit is directly CAS and the TAS from film viewer is TAS regardless the AoA. The both being same at sealevel as you noted.

gripen

[Mace2004]

Rgr that.  I agree that we should not have to worry about position/instrument error and the numbers, as presented within AH, are accurate.

Mace

[Mace2004]

Originally posted by dtango
Mace:

well since HT wrote the code he get's to define what exceeding aircraft clmax looks like so I take it from his comment that in AH physics cockpit buffet indicates that a stall has occured .

Also, I'm certain Pyro and HT are aware of the NACA report 829 that Widewing is referring to.  We've used it in the past to discuss other clmax data of particular aircraft.  If you look at that doc it discusses a variety of things including the concept of initial stall vs. the progression of the stall over the wing.  I know back when AH2 was coming out I had some conversations with Pyro that they were going to include more force vectors along the wings in AH2 which would enable finer modeling of lift across the wings.  Back then I inferred from Pyro's comments that one of the things this would allow them to do was to model stall progression across the wing among other things (e.g. increased lift due to propwash etc. etc.) since Cl doesn't just drop to 0 when Clmax is exceeded, and also that Clmax may be exceeded on certain parts of the wing prior to others.

******
HT/Pyro:

I'm curious as to what you meant by "getting behind the power curve" in relation to performing a power-on stall in AH.  I'm guessing you mean the power-required curve (drag) but not sure exactly what you were getting at in your statement.  Is there a way that we should be conducting power-on stall tests for AH planes?  I'm with Widewing that I would think power-on stalls should occur at lower airspeeds vs. power-off stalls.

Tango, XO
412th FS Braunco Mustangs

dtango,
Yeah, he's talking about being below L/Dmax.   To do level testing you need to be somewhere below L/Dmax so you can add power without accelerating, how far below is based on your specific excess power/ desired test power.  You fly level with reduced power settings and decelerate till reaching the back side of the curve.  At that point you need more power to fly slower and still stay level so you can start feeding in power until you reach test power.  You need to be stabilized at full power before the stall point is reached.  

The problem I'm seeing is in coordinating the various control inputs to counteract the increase in power during a relatively short period of time.  This is going to be highly dependant upon your specific control set up (response curves, deadbands, damping, etc.) so some may be able to do this easier than others.

[hitech]

Mace2004: Don't confuse aces high buffet with the real thing. I assume we can agree to define  stall to be the point at which  more AOA produces  less total lift?

In AH the buffet starts at that AOA .

And on behind the power curve , maces description is what I was referring to.


Widewing: The problem is in our testing (which pyro has been doing) we get the opposite results.

HiTech

[Brooke]

It's very useful to know that about the buffet in AH.  It also makes it a lot easier to do testing to have a nice clean indication of when stall is happening.

I've done some testing for various planes with power off and landing configuration.  I'll do more in clean configurations and post what I find here.  I'll also note vertical climb or descent at stall.  If it's less than about 1000 fpm, the stall speed won't be different by more than about 2% from what it would be at 0 fpm climb or descent.

I'll also post the data from the pilots' manuals (taking into account corrected indicated airspeed and correct weight of the aircraft to match the tested configuration with AH).

[Brooke]

Originally posted by Mace2004
You fly level with reduced power settings and decelerate till reaching the back side of the curve.  At that point you need more power to fly slower and still stay level so you can start feeding in power until you reach test power.  You need to be stabilized at full power before the stall point is reached.  

The problem I'm seeing is . . .

Yep, I have difficulty, too, getting into slow-flight regime.  I can manage it with some power, but I'm having trouble getting up to full power on the back side of the power curve.  It's much harder than doing it in a Cessna 152 for me.  

[Widewing]

Originally posted by hitech
Mace2004: Don't confuse aces high buffet with the real thing. I assume we can agree to define  stall to be the point at which  more AOA produces  less total lift?

In AH the buffet starts at that AOA .

And on behind the power curve , maces description is what I was referring to.


Widewing: The problem is in our testing (which pyro has been doing) we get the opposite results.

HiTech

HiTech, could Pyro e-mail a film of his procedure so that I can try to duplicate it?

My regards,

Widewing

[Stoney74]

If the buffet is the sign of the beginning of less lift with more AoA, would it be safe to assume that turn fighting into the buffet is counterproductive?  I would assume that this would mean pulling up to, but not including the buffet, would be the max AoA you should pull in a tight turn or the vertical?

[Bronk]

To me it feels like. Easing up the stick with the onset of buffet, the turn tightens.

I might be wrong though.
Widewing  would know better than I.


Bronk